X-Ray.Office

X-Ray.Office reads word-processing, spreadsheet and presentation documents and renders them to PDF and images. It is a C# port of the relevant parts of LibreOffice (Writer, Calc and Impress), developed against headless LibreOffice as the reference: a rendered page is compared with LibreOffice's rendering of the same file. It needs no LibreOffice process, office install or COM interop.

curiosity-ai/x-ray-office X-Ray.Office on NuGet

Formats

The format is always identified from the file's content, not its extension. Macro-enabled files are read as data: the package reports that macros are present and never runs them. Draw, Math and Base documents, and writing any office format, are out of scope.

Install

dotnet add package X-Ray.Office

X-Ray.Office is the facade: it identifies a file and dispatches to the right reader, and it brings the X-Ray.Office.* libraries it needs. A service that only needs one family (for example only spreadsheets) can reference that library directly instead.

Rendering uses SkiaSharp for rasterising and HarfBuzzSharp for text shaping, the shaper LibreOffice uses.

Quick start

Extract the content, without layout or fonts:

using XRay.Office;
using XRay.Office.Core.Extraction;
using XRay.Office.Markup;

ContentDocument content = OfficeDocument.Extract("report.docx");
string text = content.GetText();
string markdown = MarkdownWriter.ToMarkdown(content);

Render: open the document, lay it out (a separate, deferred step), then draw pages:

using XRay.Office;
using XRay.Office.Core.Documents;
using XRay.Office.Rendering.Pdf;
using XRay.Office.Rendering.Raster;

using IDocument doc = OfficeDocument.Open("deck.pptx");
IPageSequence pages = ((IPaginatedDocument)doc).Layout();

// One page to PNG.
using (var png = File.Create("slide1.png"))
{
    new RasterRenderer(new RasterRenderOptions { Dpi = 150 }).RenderToPng(pages[0], png);
}

// Every page to PDF.
using var pdf = File.Create("deck.pdf");
new PdfRenderer().Render(pages, pdf);

A workbook can also be written as HTML, one table per sheet, in the shape Calc's own HTML filter writes:

using XRay.Office;
using XRay.Office.Core.Documents;
using XRay.Office.Spreadsheets.Html;
using XRay.Office.Spreadsheets.Layout;

using IDocument book = OfficeDocument.Open("report.xlsx");
SpreadsheetPages sheets = (SpreadsheetPages)((IPaginatedDocument)book).Layout();
SheetHtmlWriter.Write(sheets, File.Create("report.html"));

To check what a file is without reading it, Identify reads only the container header:

using XRay.Office;
using XRay.Office.Core.Formats;

FormatIdentification format = OfficeDocument.Identify("unknown.bin");
Console.WriteLine(format.IsEncrypted);

More in this section

X-Ray.Office or X-Ray.Content?

Both read office documents. X-Ray.Content extracts text, Markdown and tables for search and LLM use across 90+ formats. X-Ray.Office lays the document out and draws it, so use it when you need pages: PDF conversion, thumbnails, previews.

License

MPL-2.0, matching the LibreOffice source it is derived from.